我正在研究嵌套的json数据,这是我使用nifi从facebook获得的。我在hive中创建了一个表,并使用该命令加载数据。
CREATE TABLE abmediaanalysis (id string, posts struct< data:array<struct< message:string, shares:struct< count:int>, id:string, reactions:struct< data:array<struct< name:string,id:string>>, paging:struct< cursors:struct< before:string,after:string>, next:string>>, likes:struct< data:array<struct< id:string>>, paging:struct< cursors:struct< before:string,after:string>, next:string>>>>, paging:struct< previous:string,next:string>>, feed struct< data:array<struct< permalink_url:string,message:string,id:string>>, paging:struct< previous:string,next:string>>)ROW FORMAT SERDE 'org.openx.data.jsonserde.JsonSerDe';
load data local inpath '/home/10879/facebook1480479682880.json' overwrite into table abmediaanalysis;
我还添加了jar文件json-serde-1.3.8-jar-with-dependencies.jar, 但当我使用横向视图爆炸打印所有列时,我得到java heap size error。我也增加了堆大小,但仍然是相同的错误
select id,posts_message,posts_share_count,posts_id,feed_data_permalink_url,feed_data_message,feed_data_id,reaction_data_name,reaction_data_id,posts_likes_data_id from abmediaanalysis
LATERAL VIEW explode(posts.data.message)MSG as posts_message
LATERAL VIEW explode(posts.data.shares.count)CT as posts_share_count
LATERAL VIEW explode(posts.data.id) I as posts_id
LATERAL VIEW explode(feed.data.permalink_url) PU as feed_data_permalink_url
LATERAL VIEW explode(feed.data.message) MSG as feed_data_message
LATERAL VIEW explode(feed.data.id) I as feed_data_id
LATERAL VIEW explode(posts.data.reactions) NM as posts_reactions_name
LATERAL VIEW explode(posts_reactions_name.data.name) NM as reaction_data_name
LATERAL VIEW explode(posts_reactions_name.data.id) NM as reaction_data_id
LATERAL VIEW explode(posts.data.likes) I as likes_data_id
LATERAL VIEW explode(likes_data_id.data.id) I as posts_likes_data_id;
当我试图打印两到三列而不是显示616条记录时,它显示约15625条记录。 谁能帮助解决这个问题 有没有机会直接从nifi加载上面的json数据到hive表?如果可以,你能告诉我。
提前致谢
答案 0 :(得分:0)
如果我可以建议您另一种方法,只需将整个JSON
字符串作为String
数据类型加载到外部表中。
e.g。
CREATE EXTERNAL TABLE json_data_table (
id String,
json_data String
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\u0001' LINES TERMINATED BY '\n' STORED AS TEXTFILE
LOCATION '/path/to/json';
使用Hive get_json_object
提取单个列。 E.g。
如果json_data列低于JSON字符串
{"store":
{"fruit":\[{"weight":8,"type":"apple"},{"weight":9,"type":"pear"}],
"bicycle":{"price":19.95,"color":"red"}
},
"email":"amy@only_for_json_udf_test.net",
"owner":"amy"
}
以下查询提取
SELECT get_json_object(json_data, '$.owner') FROM json_data_table;
返回amy
通过这种方式,您可以从表中提取每个json
元素作为列。